Audio noise determination using one or more neural networks
Apparatuses, systems, and techniques are presented to reduce noise in audio. In at least one embodiment, one or more neural networks are used to determine a noise signal in one or more speech signals.
1. A processor, comprising:
one or more circuits to use one or more neural networks to identify one or more noise signals in one or more signals based, at least in part, on two or more portions of the one or more neural networks to identify different portions of the one or more signals in parallel.
2. The processor of claim 1 , wherein the one or more circuits are further to generate an audio spectrogram corresponding to one or more features extracted from the one or more signals.
3. The processor of claim 1 , wherein the one or more circuits are further to provide an audio spectrogram as input to the one or more neural networks, wherein the one or more neural networks generate an audio mask corresponding to the noise signal identified in the one or more signals.
4. The processor of claim 2 , wherein the audio spectrogram comprises a mel spectrogram.
5. The processor of claim 1 , wherein the one or more neural networks include two parallel paths to determine patterns in an audio spectrogram, the two parallel paths including a first path comprising a sequence of convolutional layers and a second path comprising one or more gated recurrent unit (GRU) layers.
6. The processor of claim 5 , wherein the one or more circuits are further to concatenate the patterns determined by the two parallel paths and process those concatenated patterns using a sequence of GRU layers to identify important noise patterns in the one or more signals to generate the audio mask.
7. The processor of claim 6 , wherein the one or more circuits are further to invert the audio mask and apply the inverted audio mask to the audio spectrogram to generate an output audio signal with the noise signal removed from the one or more signals.
8. A system comprising:
one or more processors to use one or more neural networks to identify one or more noise signals in one or more signals based, at least in part, on two or more portions of the one or more neural networks to identify different portions of the one or more signals in parallel.
9. The system of claim 8 , wherein the one or more processors are further to generate an audio spectrogram corresponding to one or more features extracted from the one or more signals.
10. The system of claim 8 , wherein the one or more processors are further to provide an audio spectrogram as input to the one or more neural networks, wherein the one or more neural networks generate an audio mask corresponding to the noise signal identified in the one or more signals.
11. The system of claim 9 , wherein the audio spectrogram comprises a mel spectrogram.
12. The system of claim 8 , wherein the one or more neural networks include two parallel paths to determine patterns in an audio spectrogram, the two parallel paths including a first path comprising sequence of convolutional layers and a second path comprising a gated recurrent unit (GRU) layer.
13. The system of claim 12 , wherein the one or more processors are further to concatenate the patterns determined by the two parallel paths and process those concatenated patterns using a sequence of GRU layers to identify important noise patterns in the one or more signals to generate an audio mask.
14. The system of claim 13 , wherein the one or more processors are further to invert the audio mask and apply the inverted audio mask to the audio spectrogram to generate an output audio signal with the noise signal removed from the one or more signals.
15. A method comprising:
using one or more neural networks to identify one or more noise signals in one or more signals based, at least in part, on two or more portions of the one or more neural networks to identify different portions of the two or more signals in parallel.
16. The method of claim 15 , further comprising:
generating an audio spectrogram corresponding to one or more features extracted from the one or more signals.
17. The method of claim 15 , further comprising:
providing an audio spectrogram as input to the one or more neural networks, wherein the one or more neural networks generate an audio mask corresponding to the noise signal identified in the one or more signals.
18. The method of claim 15 , wherein the one or more neural networks include two parallel paths for determining patterns in an audio spectrogram, the two parallel paths including a first path comprising a sequence of convolutional layers and a second path comprising a gated recurrent unit (GRU) layer.
19. The method of claim 18 , further comprising concatenating the patterns determined by the two parallel paths and process those concatenated patterns using a sequence of GRU layers to identify important noise patterns in the one or more signals to generate the audio mask.
20. The method of claim 19 , further comprising:
inverting the audio mask and apply the inverted audio mask to the audio spectrogram in order to generate an output audio signal with the noise signal removed from the one or more signals.